ABSTRACT

Use of video is common in mobile devices. Even with high computing power and with huge amount of memory, a mobile device supports only a small number of possible video formats. A video which is unsupported needs format conversion in such a way so that it may be played at the devices at user end. This conversion process is termed as video transcoding. In this process a video is uncompressed and then again compressed with required parameters. Although devices at user end have sufficient processing power to play a video. However, the transcoding process requires even more power hence it cannot be performed at user end with current technology in real time. This transcoding can be performed in a cloud computing environment an this thesis addresses various research issues in this regard. The research work utilizes video splitting at Group of Pictures level where a video is split at Intra frames only. For simplicity, it is assumed that all group of pictures are closed group of pictures and do not have any interdependence among them. It describes overall transcoding architecture in a cloud and focuses on one type of transcoding termed as spatial resolution reduction video transcoding in which frame resolution is down scaled. The thesis also provides comparative analysis of one more mechanism termed as Temporal resolution reduction in which frame rate is reduced.